Insights on risk score development: Considerations for early-stage hepatocellular carcinoma models
Article information
Dear Editor,
We were pleased to read the paper by Ho et al. published in Clinical and Molecular Hepatology [1]. This significant study developed and validated the CATS-IF and CATS-INF risk scores to predict overall survival (OS) in patients with early-stage hepatocellular carcinoma (HCC), and we commend the authors for their rigorous efforts in this important research.
However, we would like to respectfully raise a few points for further discussion.
First, we have concerns regarding multicollinearity in the survival models. In constructing the conventional Cox regression model, the authors first performed univariate analysis to preliminarily select variables, which included albumin (ALB), albumin-bilirubin (ALBI), lymphocyte-tomonocyte ratio (LMR), and prognostic nutritional index (PNI). Subsequently, these variables were included in the multivariate analysis. However, this approach raises concerns regarding multicollinearity, as ALB, which represents serum albumin, is a component of the ALBI score, and both LMR and PNI include lymphocyte count. Given the presence of these overlapping components, we wonder whether it is necessary to assess multicollinearity issues before conducting the multivariate Cox regression analysis. Addressing this concern could enhance the reliability and interpretability of the predictive model. Additionally, in the CATS-INF model, the PNI includes both albumin and lymphocyte count, which creates multicollinearity with ALBI and LMR within the scoring system.
Second, we also wish to address concerns related to follow-up duration. The study includes patients from 2012 to 2021, with follow-up ending in June 2022. Early-stage HCC patients (BCLC 0/A) typically have a favorable prognosis, often exceeding five or even ten years of survival [2-5]. However, the relatively short follow-up for patients enrolled towards the end of the study period could result in insufficient observation time to fully assess long-term survival outcomes. This truncated follow-up may lead to high rates of right-censoring, which could introduce bias and affect the accuracy of OS estimates. To address the potential limitations associated with follow-up duration, we respectfully suggest extending the follow-up period to allow for a more accurate capture of long-term survival data. Given the gen-erally favorable prognosis for early-stage HCC patients, a minimum follow-up of five years would provide a more robust and comprehensive estimate of overall survival.
In conclusion, we greatly appreciate the authors’ efforts in developing these prognostic models and sincerely hope that addressing the concerns related to multicollinearity and follow-up duration will enhance the validity and applicability of their findings. Thank you for considering our feedback.
Notes
Authors’ contribution
Zhanna Zhang: conceptualization and manuscript draft. Gongqiang Wu: critical revision for important intellectual content, final approval.
Conflicts of Interest
The authors have no conflicts to disclose.
Acknowledgements
This work was supported by grant from the Medical and Health Science and Technology Project of Zhejiang Province (2023KY385).
Abbreviations
ALB
albumin
ALBI
albumin-bilirubin
HCC
hepatocellular carcinoma
LMR
lymphocyte-to-monocyte ratio
OS
overall survival
PNI
prognostic nutritional index